Humoral Immune Response in a Mouse Model Induced With Dengue Virus-like Particles Serotypes 1 and 4 Produced in Silkworm Larvae


 Dengue is an arboviral disease, which threatens almost half the global population, and has emerged as the most significant of current global public health challenges. In this study, we prepared dengue virus-like particles (DENV-LPs) consisting of Capsid-Premembrane-Envelope (CprM/E) and Premembrane-Envelope (prM/E) polypeptides from serotype 1 and 4, which were expressed in the silkworms using Bombyx mori nucleopolyhedrovirus (BmNPV) bacmid. 1CprME, 1prME, 4CprME, and 4prME expressed proteins in hemolymph and molecular weight of the purified proteins were 55 kDa, respectively. The purified polypeptides formed spherical Dengue virus-like particles (DENV-LPs) with approximately 30–55 nm in diameter. The immunoelectron microscopy (IEM) images revealed antigens to the surface of a lipid bilayer of DENV-LPs. The heparin-binding assay shows a positive relationship between absorbance and the quantity of E protein domain III (EDIII), which was supported by the isothermal titration calorimetry assay, showing a moderate binding affinity between heparin and DENV-LP. The high correlation between patient sera and DENV-LP reactivities revealed that these DENV-LPs shared similar epitopes with the natural dengue virus. IgG elicitation studies in mice have demonstrated that DENV-LP/CPrMEs elicits a stronger immune response than DENV-LP/prMEs, which lends credence to this claim.


Introduction
Dengue fever is a signi cant public health issue that has been reported in the Americas, Africa, Southeast Asia, Europe, the Western Paci c, and the Eastern Mediterranean. This arboviral disease is endemic in more than 100 countries, this, and approximately 96 million infected individuals have symptoms of varying severities. There has been a growing public health concern about dengue fever in the last few decades. The World Health Organization (WHO) named it one of the top ten global health threats in 2019, highlighting the critical need for a safe and effective vaccine. Despite numerous attempts, identifying the best candidate for a dengue vaccine continues to be a di cult task due to some critical factors that must be considered (Bhatt et al. 2021;Redoni et al. 2020).
Dengue virus (DENV) is the etiological agent, and it has four antigenically distinct serotypes. They belong to the Flavivirus genus and the Flaviviridae family and antigenically similar yet genetically diverse.
Based on a 70-year study that examined the spread of DENV worldwide, the most reported strains were DENV-1, DENV-2, and DENV-3, and the least frequently recorded was DENV-4. Although DENV-4 was the rst serotype of dengue to diverge in phylogenetic investigations of the Flavivirus genus, it spread slowly worldwide. DENV-1 and DENV-4 cause dengue fever with different degrees of severity. When DENV-1 infection was compared to DENV-4 infections, the duration of fever, which essentially correlates with the severity of the illness, was much longer for DENV-1. DENV-1 infection was also associated with more severe clinical manifestations than DENV-4 infection. Primary DENV-4 infection is a relatively mild sickness, but primary DENV-1 infection has more severe symptoms (Nishiura and Halstead 2007;Sang et al. 2019;Villabona-Arenas and Zanotto 2013).
Virus-like particles (VLPs) are viruses with a shell but no virus-speci c genetic material. VLPs may be able to mimic the organization and conformation of wild viruses because they contain multiprotein determinants. The organization and conformation of native viruses can be used to explore virus infection mechanisms and stimulate the host immune system to produce robust immune responses (Roldao et al. 2010). Furthermore, VLPs do not cause infections because they lack the viral genome. These properties of VLPs make them potential vaccine candidates that may be more e cient and safer than conventionally attenuated or inactivated viruses (Noad and Roy 2003).
Two successful approaches have been applied to produce recombinant VLPs of aviviruses, including capsid-premembrane-envelope (CprME) protein genes and pre-membrane-envelope (prME) proteins in cis as well as in trans from plasmid vectors. Both types of approaches should result in the formation of particles. Although the C protein is not required to form particles, the inclusion of C protein may have a stabilizing effect on VLP assembly. Capsid proteins can be arranged in one, two, or three layers, depending on their size. Some single-layer VLPs can contain more than one structural protein, whereas others cannot. When compared to the structure of single-protein VLPs (which is relatively simple), multiprotein VLPs (which contain several distinct capsid layers) have additional structural components (Krol et al. 2019;Nooraei et al. 2021).
BEVS (baculovirus expression vector system) is a high-level mass production tool for recombinant proteins in silkworm. This enables us to express eukaryotic recombinant proteins with post-translational modi cations similar to those found in mammals. The number of recombinant proteins produced by silkworm-BEVS in silkworm larvae is frequently signi cantly higher than that produced by Sf9-BEVS in cultured cells (Kato et al. 2010;Vipin Kumar Deo 2012). Feeding silkworms is exceptionally inexpensive, with a total cost of approximately USD 20 for twenty larvae. Thus, it costs slightly more than USD 20 to obtain approximately 1 mg of active Protein kinase B alpha (PKBa) (Maesaki et al. 2014). This approach is comparable in cost to an expression system based on Escherichia coli. Furthermore, because silkworms have low-cost productivity equal to the E. coli expression system, protein production can be quickly and inexpensively scaled up. This expression system is favorable for vaccine development (Fujita et al. 2020).

Western blot analysis
To separate the proteins, 10% or 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) was used and subsequently subjected to western blotting by blotting the proteins trapped in acrylamide gel onto an Immobilon-P PVDF (polyvinylidene uoride) membrane (Merck Japan) using the Mini Trans-Blot Electrophoretic Transfer Cell (Bio-Rad, Hercules, CA, USA). After blotting, the membrane -3 was blocked in Tris-buffered saline with 0.1% Tween 20 detergent (TBST) (pH 7.6) with 5% skim milk (FUJIFILM Wako Pure Chemical) and then incubated in 10000-fold diluted mouse anti-HA tag antibody (Medical & Biological Laboratories) for DENV-1 constructs and anti-strap tag II antibody (Medical & Biological Laboratories) for DENV-4 constructs. Alternatively, 1500-fold diluted speci c serotype monoclonal anti-envelope antibodies, anti-E DENV-1 E29 clone (BEI Resources, Virginia, US) for DENV-1 constructs, and anti-E DENV-4 E42 clone (BEI Resources) for DENV-4 constructs were used as the primary antibodies. Anti-E DENV-2 3H5-1 clone (BEI Resources) and anti-E DENV-3 E1 clone (BEI Resources) antibodies were also used to assess cross-reactivity between serotypes for all puri ed proteins. After three washes with TBST, the membrane was incubated for one hour with 10000-fold diluted anti-mouse IgG antibody conjugated to horseradish peroxidase (HRP) (FUJIFILM Wako Pure Chemical). The speci c bands were discovered using a Fluor-S MAX Multi-Imager (Bio-Rad).

Heparin-binding assay by ELISA
The heparin-binding assay by ELISA was performed with modi cations as previously described (Utomo et al. 2019;Utomo et al. 2020). Six ng/ml diluted biotin-labeled heparin (Sigma-Aldrich Japan) and 1.8 ng of heparin were immobilized into blockless avidin plate (Sumitomo Bakelite, Tokyo, Japan) wells and washed three times with PBS. For a negative control, 2 μg of BSA was used. Puri ed proteins at various concentrations (0.5, 1, 5, and 10 g/ml) were added to wells at corresponding quantities, incubated at room temperature for one hour, and then washed with PBST. After serial washing, a 1000-fold diluted rabbit anti-DENV E polyclonal antibody (GeneTex) was added, followed by a 1000-fold diluted HRPconjugated anti-rabbit IgG antibody (FUJIFILM Wako Pure Chemical). To stop the reaction, 100 μl of substrate 0.1 mg/ml 3.3',5.5'-tetramethylbenzidine (TMB) in 100 mM sodium acetate (CH 3 COONa), pH 6.0, was added to each well with 0.2% (v/v) 30% hydrogen peroxide, and 50 μl of 1 N H 2 SO 4 was added.
The absorbance was estimated at 450 nm.
Isothermal Titration Calorimetry assay for heparin binding to DENV-LPs The binding a nities of DENV-LP to heparin were determined using isothermal titration calorimetry (ITC) on a MicroCal iTC200 (Malvern Panalytical Ltd, Enigma Business Park, UK). Titrations were performed at a temperature of 25 °C by injecting 2 μl aliquots of 1000 μM ligand dissolved in 1× PBS buffer into a cell containing 10 μM DENV-LP. The heat release was recorded, and the titration data were analyzed with MicroCal Origin ITC software (Malvern Panalytical Ltd). Thermodynamic parameters were determined by tting experimental data with nonlinear least-squares using the one-set sites binding model (Duff et al. 2011;Ikegaya et al. 2021).

Human DENV-infected sera interaction with DENV-LPs
Direct ELISA was carried out with modi cations as previously described (Utomo et al. 2019;Utomo et al. 2020). An interaction between antigens, 1CprME, 1prME, 4CprME, and 4prME, and patient sera was detected using a direct ELISA method. Dengue patient sera [rapid diagnostic test NS1(+)] were used. Sera were obtained from the Centre of Pharmaceutical and Medical Technology, National Research and Innovation Agency, Jakarta, Indonesia. Protocols of the collection were reviewed and approved by the Health Research Ethics Committee-University of Indonesia and Cipto Mangunkusumo Hospital (HREC-FMUI/CMH) (approval no. KET-1358/UN2.F1/ETIK/PPM.00.02/2020).
Each diluted sample, 100 µl of 20 ng/ml in coating buffer (0.05 M carbonate-bicarbonate, pH 9.6), was applied to a 96-well ELISA plate and incubated overnight at 4 °C. After incubation, the coating solution was discarded, and a 100 µl blocking solution (5% skim milk in PBS) was added to each well, followed by 1 h at 37 °C incubation. The plates were then washed serially with PBST buffer before adding 100 µl of 1:50 patient sera in PBS. Plates were then incubated at 37 °C for 1 h before being washed three times with washing buffer and before the addition of 100 µl of 1:5000 anti-human IgG-HRP conjugated antibody. Plates were then incubated at 37 °C for 1 h and washed, and 50 µl of TMB substrate was applied and incubated for 10 min before being stopped with 50 µl of 1 M H 2 SO 4 . The absorbance was measured at 450 nm.

Immunoelectron microscopy
Immunoelectron microscopy (IEM) was carried out as previously described (Utomo et al. 2020) with modi cations. The puri ed antigen sample was added to the Cu-grid transmission electron microscopy (TEM) (Nisshin EM Co., Ltd., Tokyo) and incubated for 30 sec at room temperature, washed with 30 µl of PBS, and incubated for 30 sec, repeated three times. BSA (30 µl of 2% v/v) was used for blocking after adding a puri ed antigen sample and washed three times with PBS. The Cu-grid was washed in stages. The rst and secondary antibodies were anti-E rabbit polyclonal antibody (FUJIFILM Wako Pure Chemical) diluted 30 times and goat anti-rabbit IgG-conjugated (FUJIFILM Wako Pure Chemical) gold nanoparticles diluted 50 times, respectively. The Cu grid was treated with 1% phosphotungstic acid and analyzed with a JEM-2100F TEM system (JEOL Ltd., Tokyo, Japan).

Mice immunization
Mouse immunization was carried out as previously described (Utomo et al. 2020) with modi cations. In this study, 50 BALB/c mice aged 4-6 weeks were divided into ten groups: i) negative control (PBS), ii) immunized with alhydrogel as an adjuvant, iii) immunized with 1CprME, iv) immunized with 1CprME+adjuvant, v) immunized with 1prME, vi) immunized with 1prME+adjuvant, vii) immunized with 4CprME, viii) immunized with 4CprME+adjuvant, ix) immunized with 4prME, and x) immunized with 4prME+adjuvant. All mice were kept in a temperature-controlled, light-cycled room and were divided into ten groups based on the immunogen. Each mouse was immunized three times intraperitoneally within two weeks with 50 mg of puri ed 1CprME, 1prME, 4CprME, and 4prME proteins with alhydrogel adjuvant. Blood samples were collected via the tail vein on Days 0, 16, and 30, and sera were isolated and stored at -80 °C. All animal procedures were conducted in compliance with established guidelines from the Animal Laboratory of Center of Pharmaceutical and Medical Technology, National Research and Innovation Agency, Indonesia. Animal experiment protocols were reviewed and approved by the Health Research Ethics Committee-University of Indonesia and Cipto Mangunkusumo Hospital (HREC-FMUI/CMH) (approval no. KET-721/UN2.F1/ETIK/PPM.00.02/2021).

Results
Expression of 1CprME, 1prME, 4CprME and 4prME polypeptides in silkworm DENV structural proteins are composed of a C and two membrane proteins, prM and E, translated at the beginning of the polyprotein in the order C-prM-E. BmNPV/1CprME (Fig. 1a), BmNPV/1prME (Fig. 1b), BmNPV/4CprME (Fig. 1c), and BmNPV/4prME (Fig. 1d) bacmids were injected into silkworm larvae, and the silkworm hemolymph was collected at 5 dpi. The expression of 1CprME, 1prME, 4CprME, and 4prME in hemolymph samples was con rmed, with molecular weights of 55 kDa at the E protein (Fig. S1 of Supplementary Information), which corresponded to the estimated weight.
Puri cation of 1CprME, 1prME, 4CprME and 4prME polypeptides The 1CprME, 1prME, 4CprME, and 4prME polypeptides were puri ed by a nity chromatography and con rmed by western blot using mouse anti-HA tag antibody for DENV-1 constructs (Fig. 2a, b) and antistrap tag II antibody for DENV-4 constructs, which showed bands of size 55 kDa at elution fractions (Fr), respectively. To determine whether the puri ed 1CprME, 1prME, 4CprME, and 4prME polypeptides contained E proteins, western blotting was performed using serotype-speci c monoclonal anti-envelope antibodies, an anti-E DENV-1 E29 clone for DENV-1 constructs, and an anti-E DENV-4 E42 clone for DENV-4 constructs. All of the constructs' bands were con rmed to be approximately 55 kDa (Fig. 2e, f). These results demonstrate that the E proteins were present in the puri ed 1CprME, 1prME, 4CprME, and 4prME polypeptides. Anti-E-DENV-1 E29 clone, Anti-E-DENV-2 3H5-1 clone, Anti-E-DENV-3 E1 clone, and Anti-E-DENV-4 E42 clone antibodies were used to investigate the cross-reactivity of all puri ed proteins. Speci c bands for 1CprME (Fig. 3a), 1prME (Fig. 3b), 4CprME (Fig. 3c), and 4prME ( Fig. 3d) could not be detected using the speci c serotype antibodies. These results indicate that there is no cross-reactivity between speci c serotype antibodies and all of the constructs.
Morphology of 1CprME, 1prME, 4CprME and 4prME polypeptides IEM was used to con rm the morphology of the polypeptides 1CprME, 1prME, 4CprME, and 4prME. Spherical structures with sizes ranging from 30 to 55 nm were observed (Fig. 4a-d), supported by data from dynamic light scattering (Fig. 4e-h). The IEM observation revealed that the particles were lipid bilayer-structured spherical, with some immunogold bound to their surface. The presence of anti-dengue E protein in gold nanoparticles bound to the surface of spherical structures indicates that the particles contain dengue E protein on the surface of the VLPs. According to these results, the 1CprME, 1prME, 4CprME, and 4prME polypeptides expressed in silkworms are capable of generating VLPs of Dengue-1 and -4 (DENV-LPs/1CprME, DENV-LPs/1prME, DENV-LPs/4CprME, and DENV-LPs/4prME).
Heparin-binding assay of the DENV-LPs/1CprME, /1prME, /4CprME, and /4prME To con rm the expression of EDIII on the surface of DENV-LPs, a heparin-binding assay was performed. The binding assay of the puri ed DENV-LPs/1CprME, /1prME, /4CprME, and /4prME was performed using heparin-immobilized microtiter plates. In ELISA, the absorbance increased proportionally to the presence of E (Fig. 5a). These results indicate that the EDIII domain is present on the surface of the DENV-LPs of 1CprME, 1prME, 4CprME, and 4prME.

Isothermal Titration Calorimetry assay for heparin-binding to DENV-LPs
Binding activities of DENV-LPs/1CprME, /1prME, /4CprME, and /4prME toward heparin were investigated by ITC (Fig. 5b-e). Heparin binds to DENV-LPs with K D values of 21-51 mM and ΔG values from -5.5 to -6.2 kcal mol -1 (Table 2). K D values, 29.5 mM of DENV-LPs/1CPrME or 21.1 mM of /4CprME were 1.7-fold lower compared to 51.0 mM of DENV-LPs/1prME or 36.2 mM of /4prME, suggesting CprMEs are higher a nity to heparin than prMEs which are the similar results of ELISA (Fig. 5a). ΔG value of DENV-LPs/1CprME or /4CprME was lower than that of DENV-LPs/1PrME or /4prME, indicating that the heparin binding to the EDIII on CprMEs are more spontaneous than to prMEs. Table 2 The binding a nity of DENV-LPs towards heparin.
The interaction between heparin and DENV-LPs has been analyzed using multiple assays, including ELISA with immobilized heparin and ITC. Heparin binds to DENV-LPs with K D values of 21-51 mM and ΔG values from -5.5 to -6.2 kcal mol -1 using the ITC system. According to other references, Marks et al. (2001) used O-sulfated heparin or Nand O-sulfated heparin to bind to the DENV-2 envelope protein expressed in E. coli and analyzed using SPR, yielding a K D of 5 nM. Chen et al. (1997) used heparin to analyze the binding to the DENV-2 envelope protein expressed on COS-7 mammalian cells. They determined the K D to be 15 nM using ITC. Kim et al. (2017) used heparin to analyze the binding of the ZIKV envelope protein expressed in E. coli and determined the K D to be 433 nM using SPR (Table 2). In comparison to a protein subunit, VLPs are a complex molecule composed of several protein subunits. It is known that the a nity of the ligand-protein is based on the surface's interaction, leading to the conjugation within the pocket of the protein. Hence, according to the structure of the VLP, it is considered relatively high molar mass of macromolecules which affect the equilibrium and kinetics of its protein function. Consequently, the low effective molar concentration of a VLP results to low surface available for binding and result to overall a nity of a VLP binding being underestimated. Regarding antiviral activity, sulfated heparin has better binding ability than heparin. Therefore, sulfated heparin is used to inhibit viral infection of cells. However, unmodi ed heparin in the cell membrane is known to bind and interact with EDIII, the putative receptor-binding domain in the avivirus E protein crystal structure (Frei et al. 2018;Han et al. 2018;Yang et al. 2016;Zautner et al. 2006). EDIII also contains epitopes that block viral adsorption and are targeted by many antibodies, including serotype-speci c neutralizing monoclonal antibodies. The conformational exibility of heparin might permit this molecule to more easily adopt a productive conformation for interaction with the envelope protein.
Although it has only a moderate binding a nity, the envelope of DENV-LP is still recognized by heparin on the surface of the cell (Hyatt et al. 2020;Marks et al. 2001). While mammalian cells secrete proteins with correct conformations and full biological activity, insect cells offer advantages comparable to those of mammalian cells when compared to E. coli or yeast expression systems. The ability to introduce foreign DNA into these cells facilitates a better understanding of the transcriptional, translational, and posttranslational machinery in mammalian cells (Gray 2001;Ikonomou et al. 2003).
The antigens of 1CprME, 1prME, 4CprME, and 4prME also showed reactivity to patient sera high a nity in the direct ELISA. Mixed sera of dengue patients can react to many types of nonspeci c DENV epitopes. The strongly correlated reactivities of patient sera with DENV-LPs indicated that the same epitope(s) were displayed on these DENV-LPs. However, there will be a marked difference in the level of antibody reactivity. In other words, the epitope(s) displayed on the DENV-LP surface is comparable with those of the native dengue virus (Danko et al. 2018;Wang et al. 2003).
DENV-LPs/1CprME, /1prME, /4CprME, and /4prME could generate IgG antibody levels since many VLPs include structural or molecular characteristics that give certain auto-immunostimulatory qualities. These features enable VLPs to induce immunological responses without the use of adjuvants; nevertheless, adjuvants may enhance vaccination immunogenicity and encourage the activation of a speci c type of immune response when combined with VLP vaccines (Cimica and Galarza 2017;Donaldson et al. 2018;Müller et al. 2020). Our ITC data indicate that DENV-LP/1CPrME and DENV-LPs/4CPrME have a lower K D and ΔG than DENV-LP/1prME and DENV-LPs/4prME. This demonstrates that the CPrME constructs have a higher a nity and more spontaneous binding then the prME constructs. Those reactions can be harnessed to perform work inside the body. The IgG elicitation in mice con rmed DENV-LP/CPrMEs elicit a stronger immune response than DENV-LPs/prMEs, indicates CprME can easily bind to heparin-like receptor on the surface of the cells compare to prME. When mice were immunized with avivirus VLPs, CprME VLPs exhibited superior antigenicity to prME VLPs. Due to the superiority of the CprME VLP, a capsid should be included in the vaccine to improve immunity. This repetitive protein structure can boost innate immunity and prompt B cells to directly generate neutralizing antibodies (Garg et al. 2019;Nooraei et al. 2021). Each DENV serotype carries the conserved antibody epitope incorporated in the N-and C-terminal regions of the C protein and is e ciently recognized by dengue patients exposed to primary and secondary infections from other serotypes. The C-protein central region has an epitope of the peptide, primarily targeted by serotype-speci c antibodies (Alves et al. 2016;Nadugala et al. 2017;Rana et al. 2018 international, national, and/or institutional guidelines for the care and use of animals were followed.

Consent for publication
Not applicable.

Availability of data and material
All the data and materials have been provided in the main manuscript.

Competing interests
The authors declare that they have no competing interests.

Funding information
This work has been funded by the Japan Society for the Promotion of Science (JSPS) KAKENHI Grant-in-Aid for Scienti c Research (A) (Grant No. 20H00411) and partly (16H02544) and partly by the Heiwa Nakajima Foundation's Asian region priority academic research grant.
Authors' contribution DISU was the main researcher for this study and was involved with the experimental design and operation. SP participated in animal experiments and provided resources, respectively. EYP provided ideas, funded the research, revised the manuscript, and supervised this study. All authors read and approved the manuscript.

Figure 1
Construction of recombinant dengue virus structural proteins expressed in this study. (a) 1CprME, (b) 1prME, (c) 4CprME, and (d) 4prME polypeptides of DENV-1 and 4 were expressed in silkworms as a fusion protein with HA+FLAG tags for DENV-1 and Strep-tag II+FLAG tags for DENV-4.

Figure 2
Western blot of puri ed (a) 1CprME, (b) 1prME, (c) 4CprME and (d) 4prME polypeptides. Each protein was puri ed from silkworm hemolymph using anti-FLAG tag protein puri cation gel column chromatography.
E protein was veri ed using speci c serotype monoclonal antibodies using e an anti-E E29 clone for DENV-1 constructs and f an anti-E E42 clone for DENV-4 antibodies.

Figure 3
Western blot of puri ed (a) 1CprME, (b) 1prME, (c) 4CprME, and (d) 4prME polypeptides for the crossreaction test was performed on the DENV-1 and DENV-4 constructs with speci c serotype monoclonal antibodies for each serotype.